Denoising Image of Mammography Using Wavelet Methods

Abstract

Breast cancer is a tumor formed by breast cells that grow and develop uncontrollably. Detection of breast cancer can be done with the results of mammography images. In mammography images, there is often noise and a gray level. This study aimed to improve the quality of mammographic images by reducing noise and gray levels from the image to examine cancer areas. This research was conducted by improving the quality of mammographic images using the Haar, Daubechies, Biorthogonal, and Symlet wavelet methods. The tissue classification used is circumscribed, asymmetry, calcification, and normal. Testing the results of improving the quality of mammography images needs to be done using the calculation of the value of MSE and PSNR. In the denoising results of a circumscribed image using the haar wavelet method, the MSE results are 0.0002, and the PSNR results are 37.8425 dB in the salt&pepper noise image, while in the quantum noise image the MSE results are 0.0008 and PSNR results are 31.0751 dB. The results of this study show that the Haar wavelet method can be used for noise reduction in salt and pepper images. In contrast, the Haar, Daubechies, Biorthogonal, and Symlet wavelet methods can be used on images with quantum noise.